#1 • AI Assistant
What it is: Software you can talk to that helps you think or write.
Why you’d care: It reduces mental load.
Example: Asking it to clean up an email or explain something simply.
AI & Tech Terms
Plain-English index — what it is, why it matters, example.
How to use this index
- Look up any term you hear or see
- Read “What it is” if short on time
- Bookmark and return anytime
#2 • LLM (Large Language Model)
What it is: The engine behind AI assistants.
Why you’d care: Better engines usually give better answers.
Example: The hidden system powering chat-based AI tools.
#3 • Prompt
What it is: The instruction or input you give an AI system.
Why you’d care: Clear prompts lead to clearer results.
Example: Asking an AI to summarize something in plain English.
#4 • AI Agent
What it is: An AI system designed to take actions, not just respond.
Why you’d care: Agents move AI from conversation into execution.
Example: An agent that schedules meetings automatically.
#5 • SLM (Small Language Model)
What it is: A smaller model trained for specific tasks.
Why you’d care: SLMs are faster and easier to control.
Example: A private model trained only on company data.
#6 • VPS (Virtual Private Server)
What it is: A rented server you control on the internet.
Why you’d care: It provides independence from platforms.
Example: Hosting your own site or AI tools.
#7 • API
What it is: A way for software systems to communicate.
Why you’d care: APIs connect tools and automate work.
Example: Sending text to an AI model and receiving a response.
#8 • Token
What it is: A small chunk of text processed by an AI.
Why you’d care: AI reads tokens, not words.
Example: A sentence split into many tokens.
#9 • Context Window
What it is: How much information an AI can remember at once.
Why you’d care: Older info drops out when the window fills.
Example: Long chats causing earlier details to be forgotten.
#10 • Weights
What it is: Internal values that shape AI responses.
Why you’d care: They determine what patterns are recognized.
Example: Different weights create different behaviors.
#11 • Training Data
What it is: Information used to teach an AI model.
Why you’d care: Models reflect what they were trained on.
Example: Books and articles used during training.
#12 • Inference
What it is: When an AI generates an output.
Why you’d care: Every response is an inference.
Example: Typing a question and receiving an answer.
#13 • Fine-Tuning
What it is: Adjusting a model for a specific purpose.
Why you’d care: Improves results for narrow tasks.
Example: Training a model on support emails.
#14 • Self-Hosted
What it is: Software running on servers you control.
Why you’d care: Increases ownership and flexibility.
Example: Running tools on your own VPS.
#15 • Platform Lock-In
What it is: Difficulty leaving a service once dependent.
Why you’d care: Limits future options.
Example: Workflows tied to a single platform.
#16 • Automation
What it is: Tasks executed automatically by systems.
Why you’d care: Saves time when designed correctly.
Example: Auto-sending follow-up emails.
#17 • Human-in-the-Loop
What it is: AI systems supervised by humans.
Why you’d care: Prevents costly mistakes.
Example: Reviewing AI output before publishing.
#18 • Neural Network
What it is: A layered system that processes information.
Why you’d care: Most modern AI relies on neural networks.
Example: Text flowing through model layers.
#19 • Neuron
What it is: A single processing unit in a neural network.
Why you’d care: Neurons are AI’s basic building blocks.
Example: Millions activating per response.
#20 • Model
What it is: A trained AI system that produces outputs.
Why you’d care: Defines what the AI can do.
Example: A language model generating text.
#21 • Dataset
What it is: A collection of data used to train or test a model.
Why you’d care: Data quality shapes model behavior.
Example: Documents used to teach an AI writing patterns.
#22 • Overfitting
What it is: When a model learns training data too closely.
Why you’d care: It fails on new inputs.
Example: Memorizing examples instead of generalizing.
#23 • Generalization
What it is: A model’s ability to handle new situations.
Why you’d care: Makes AI useful beyond training.
Example: Answering unseen questions.
#24 • Hallucination
What it is: AI producing confident but incorrect output.
Why you’d care: Requires human verification.
Example: Inventing facts or citations.
#25 • Latency
What it is: The delay between request and response.
Why you’d care: High latency feels slow.
Example: Waiting seconds for a reply.
#26 • Throughput
What it is: How much work a system can handle at once.
Why you’d care: Matters when scaling.
Example: Processing many requests per minute.
#27 • DNS
What it is: System that maps names to IP addresses.
Why you’d care: Makes the internet usable.
Example: Typing a domain instead of numbers.
#28 • Domain
What it is: A human-readable internet address.
Why you’d care: Domains are digital real estate.
Example: Owning your own domain name.
#29 • SSL
What it is: Encryption securing data in transit.
Why you’d care: Protects privacy and trust.
Example: The lock icon in the browser.
#30 • Firewall
What it is: A system that controls network traffic.
Why you’d care: Reduces exposure to attacks.
Example: Blocking unauthorized access.
#31 • Port
What it is: A numbered communication endpoint on a server.
Why you’d care: Ports determine how services connect.
Example: HTTPS traffic commonly uses port 443.
#32 • Container
What it is: A packaged environment that runs software consistently.
Why you’d care: Containers reduce setup and compatibility issues.
Example: Running an app the same way everywhere.
#33 • Stack
What it is: A collection of tools used together.
Why you’d care: Stacks shape complexity and maintenance.
Example: Server + database + frontend.
#34 • Subscription
What it is: Recurring payment for continued access.
Why you’d care: Subscriptions quietly accumulate.
Example: Monthly SaaS fees.
#35 • SaaS
What it is: Software delivered over the internet.
Why you’d care: Convenient but limits control.
Example: Web-based tools.
#36 • Workflow
What it is: A defined sequence of steps.
Why you’d care: Prevents automation chaos.
Example: Lead → follow-up → archive.
#37 • Integration
What it is: Connecting systems so they work together.
Why you’d care: Adds power but adds fragility.
Example: Form connected to CRM.
#38 • Integration Debt
What it is: Complexity from too many connections.
Why you’d care: Makes changes risky over time.
Example: Breaking automations after updates.
#39 • Ownership
What it is: Control over assets and systems.
Why you’d care: Ownership creates leverage.
Example: Owning your site and data.
#40 • Access
What it is: Permission without ownership.
Why you’d care: Access can be revoked.
Example: Using a platform account.
#41 • Context
What it is: Information an AI uses to understand what you’re asking.
Why you’d care: Better context produces better answers.
Example: Giving background before a question.
#42 • System Prompt
What it is: Instructions that guide how an AI behaves.
Why you’d care: They shape tone and boundaries.
Example: Telling an AI to explain things plainly.
#43 • Temperature
What it is: Controls how predictable or creative responses are.
Why you’d care: Lower is factual, higher is creative.
Example: Low temperature for summaries.
#44 • Vector
What it is: A numeric representation of meaning.
Why you’d care: Allows similarity search.
Example: Finding related documents.
#45 • Embedding
What it is: Text converted into vectors.
Why you’d care: Enables semantic retrieval.
Example: Searching notes by meaning.
#46 • RAG (Retrieval-Augmented Generation)
What it is: AI that pulls from external data before answering.
Why you’d care: Improves accuracy.
Example: Answering using your documents.
#47 • Rate Limit
What it is: A cap on how often a system can be used.
Why you’d care: Prevents abuse and overload.
Example: API requests per minute.
#48 • Agent Loop
What it is: A cycle of plan, act, and review.
Why you’d care: Lets agents work step by step.
Example: Plan → act → check → repeat.
#49 • Cyborg
What it is: A human augmented by technology.
Why you’d care: Modern work blends human + AI.
Example: Using AI as a thinking partner.
#50 • Inference Cost
What it is: The cost of running AI outputs.
Why you’d care: Matters when scaling usage.
Example: Paying per request.
Updated January 2026